HumanCompatibleAI / evaluating-rewards

Library to compare and evaluate reward functions
https://arxiv.org/abs/2006.13900
Apache License 2.0
61 stars 7 forks source link

Heatmaps of reward for illustrative gridworlds #10

Closed AdamGleave closed 4 years ago

AdamGleave commented 4 years ago

Tasks:

codecov[bot] commented 4 years ago

Codecov Report

Merging #10 into master will increase coverage by 0.06%. The diff coverage is 82.25%.

Impacted file tree graph

@@            Coverage Diff             @@
##           master      #10      +/-   ##
==========================================
+ Coverage   81.35%   81.42%   +0.06%     
==========================================
  Files          41       43       +2     
  Lines        2499     2745     +246     
==========================================
+ Hits         2033     2235     +202     
- Misses        466      510      +44
Impacted Files Coverage Δ
src/evaluating_rewards/analysis/plot_pm_reward.py 87.8% <ø> (ø) :arrow_up:
...rc/evaluating_rewards/scripts/train_preferences.py 97.82% <ø> (ø) :arrow_up:
src/evaluating_rewards/scripts/train_regress.py 95.12% <ø> (ø) :arrow_up:
...uating_rewards/analysis/plot_divergence_heatmap.py 70.83% <ø> (ø) :arrow_up:
src/evaluating_rewards/analysis/stylesheets.py 71.42% <ø> (ø) :arrow_up:
src/evaluating_rewards/scripts/model_comparison.py 73.17% <ø> (ø) :arrow_up:
tests/test_scripts.py 100% <100%> (ø) :arrow_up:
src/evaluating_rewards/analysis/visualize.py 83.24% <100%> (+0.09%) :arrow_up:
...luating_rewards/analysis/plot_gridworld_heatmap.py 53.93% <53.93%> (ø)
...c/evaluating_rewards/analysis/gridworld_heatmap.py 98.05% <98.05%> (ø)
... and 2 more

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 24d90b2...52972be. Read the comment docs.